The Most Reliable Hard Drive?
Which hard drive brands are the most reliable? And what is the likelihood that your hard drive is going to crash and burn in the near future? Read on to learn more about hard drive reliability and diagnostic tools...
Which Hard Drives Last Longest?
A hard drive failure is the worst thing that can happen to your computer. If any other component fails you can replace it and get back to work pretty quickly. But when a hard drive stops working you lose access to all of your personal data stored on it. Getting that data back can be time-consuming, expensive, or even impossible, especially if you haven’t kept adequate backup copies. So the question, “What is the most reliable hard drive?” is particularly pertinent.
A massive amount of data that can help to answer that question is coming out of cloud-storage provider BackBlaze these days. Backblaze is unique among cloud services in running an “open source” operation; it regularly publishes detailed blog posts about the composition and performance of its infrastructure.
The company uses consumer-grade hard drives, most of 3 or 4 Terabyte capacities, so their performance is relevant to the average home or small business user. As of December 31, 2014, Backblaze had 41,213 disk drives in service, a more than adequate sample size. Here’s the gist of their Q4 2014 reliability report, below. (See also, the updated Q3 2015 reliability report.)
The Seagate Desktop HDD.15 4-Terabyte drive ($140 at Newegg) provided the best price/reliability ratio, with an annual failure rate of 2.6%. Backblaze has over 12,000 of these drives; their average age is 0.9 years.
The HGST (formerly Hitachi) Deskstar 4000 4-Terabyte drive ($175 at Amazon) costs a bit more but had an annual failure rate of 1.4%. Backblaze has about 12,000 of these drives, too, with an average age of 0.4 years.
Yes, these are very young drives on average. The annual failure rates of “middle-aged” (2-4 years old) drives from Seagate and HGST are even lower. Failure rates increase dramatically as drive age exceeds 4 years, e.g., 9.5% and even 23.5%. This pattern is easily explained.
There are three general categories of drive failure causes. The first is “infant mortality,” or manufacturing defects which cause drives to fail quickly after they are put into use. The second source of failures is “random events” such as physical shocks, power blips, heat damage, cosmic rays, and so on; these events may occur at any time. The third is “old age” -- everything wears out eventually. Add them all together and you get a “bathtub curve” plot of failure rates vs. average age: high at both ends, relatively low in the middle.
My takeaway from this report was that the Hitachi Deskstar (3TB and 4TB) models seem to offer the best reliability of all consumer-grade hard drives. Also, the Seagate 4TB models look like better long-term performers than the 1.5 and 3TB offerings.
Indications of Impending Failure
Backblaze’s failure rate stats are interesting if you are shopping for a new drive, but they don’t really tell you if it’s time to start shopping. That is, they don’t say anything about whether your current hard drive is about to fail. However, another Backblaze blog post is highly pertinent to that important question!
Every modern hard drive has SMART (Self-Monitoring, Analysis, and Reporting Technology) built into it. SMART enables monitoring of 70 metrics of a hard drive’s operation. Backblaze has identified five SMART metrics that strongly correlate with impending drive failure; they are:
- SMART 005 (05) - Reallocated_Sector_Count.
- SMART 187 (BB) - Reported_Uncorrectable_Errors.
- SMART 188 (BC) - Command_Timeout.
- SMART 197 (C5) - Current_Pending_Sector_Count.
- SMART 198 (C6) - Offline_Uncorrectable
Some of these SMART metrics may not be implemented by a drive manufacturer. My Toshiba drive does not report SMART 187 & 188, for example. To check these and other SMART metrics, I recommend the Passmark DiskCheckup utility; it’s free for personal use or only $19 for commercial use. DiskCheckup and similar utilities use hexadecimal notation for the SMART metrics; hex values are given in parentheses above.
The results of the DiskCheckup tool can be somewhat cryptic, as you see in the sample output from one of my own hard drives. The status column shows that everything is "OK" but some of those numbers in the "Raw Value" column are a bit worrying. I'm more inclined to give credence to the Raw Value numbers, and run some additional diagnostics on this drive. (See links to free diagnostic tools in my article Is Your Hard Drive Going to Crash?)
The good news is that even if my hard drive suffers a sudden failure today, I won't lose any work. Why? Because I do weekly full image backups, supplemented by nighly incremental backups and "real-time" backups of my most important folders. If you're confused by any of that, or concerned that your data isn't properly backed up, I recommend my ebook Everything You Need to Know About Backups.
Your thoughts on this topic are welcome. Post your comment or question below...
This article was posted by Bob Rankin on 26 Feb 2015
|For Fun: Buy Bob a Snickers.|
Geekly Update - 25 February 2015
The Top Twenty
Computer Backup Power - What You Need to Know
Post your Comments, Questions or Suggestions
Free Tech Support -- Ask Bob Rankin
Subscribe to AskBobRankin Updates: Free Newsletter
Copyright © 2005
- Bob Rankin - All Rights Reserved
Article information: AskBobRankin -- The Most Reliable Hard Drive? (Posted: 26 Feb 2015)
Copyright © 2005 - Bob Rankin - All Rights Reserved